ASTRID vs S88

The ASTRID methodology has been presented in MESURE recently (January
1999). One may remind of a few features to clarify what are S88 and
ASTRID for each other:

The terminology does not match. It is a shame: a common language would
help the mutual understanding of people. However, the compliance of
ASTRID terminology to S88 could be easily obtained.

As far as the implementation is concerned, the S88.01 standard is much
less directive than ASTRID: before the release of S88.02, it is still
difficult to state than a system is S88 compliant or not. The
availability of JMR's development tools (X-Nodes) and the fully
controlled spread of the method had led to promote ASTRID as a quite
complete " de facto " standard in the users' community.

The ASTRID procedural hierarchy model defines only one procedural
element, the ``Function''. The Recipe object in charge of the Function
activation is not detailed. 

The S88 procedural hierarchy model is sophisticated. The procedural
objects are distributed on 3 levels and may be freely dispatched between
the Recipe (managed by the process engineer) and the Equipment (managed
by the process control engineer). These levels can be expanded or
collapsed according the control strategy 

The S88 standard defines several Recipe types, from the General Recipe
developed by the R&D lab to the Master and Control Recipe used at the
shop floor. However, even the second part of the standard does not
explain how to manage the consistency between these recipes. Focused on
the product conception process, the S88 standard ignores the other
executive business processes (modelling for planning and scheduling for
instance). 

The ASTRID methodology defines a framework for the development of its
functional objects but does not define a language to describe the
procedures while S88.02 standard will propose a graphical language. 

ASTRID presents some particular features, which can lead into problems
then trying to implement the method: split between the behaviour and the
interface of the function, ergonomics of HMI...

To the credit of similarities, we notice that:

" Function ", the ASTRID upper (and only) procedural element may
correspond to the " Phase ", the S88 lowest procedural element

ASTRID " Resource " and " Device " are similar to S88 " Equipment Module
" and " Control Module ". However, the S88 standard put these element in
a different modelling space

Thus ASTRID defines a possible implementation for equipment control
while S88 only states principles. Even the second part does not give any
guidelines, considering that this subject rely on implementation domain.

In conclusion, the S88 standard never conflicts with ASTRID, which can
be used to implement an S88 compliant system.

ASTRID does not tackle the same issues than the standard. It mainly
concerns with the robust control of equipments while the standard
concerns with physical, process and procedural models of batch control. 

It is reassuring to notice the clearsightedness of the S88 designers and
the pragmatism of the ASTRID promoters who join on the same boundary. 

In fact, it is less than possible that the equipment control will be
standardized one day: a lot of solutions similar to ASTRID are proposed
by vendors or developed by users for ages. One never thought about
standardizing the process control system software development...